The effects of high-intensity interval training vs. moderate-intensity continuous training on body composition in overweight and obese adults: a systematic review and meta-analysis: Exercise for improving body composition

Abstract

Objective:
The objective of this study is to compare the effects of high-intensity interval training (HIIT) and moderate-intensity continuous training (MICT) for improvements in body composition in overweight and obese adults.

Methods:
Trials comparing HIIT and MICT in overweight or obese participants aged 18-45 years were included. Direct measures (e.g. whole-body fat mass) and indirect measures (e.g. waist circumference) were examined.

Results:
From 1,334 articles initially screened, 13 were included. Studies averaged 10 weeks × 3 sessions per week training. Both HIIT and MICT elicited significant (p < 0.05) reductions in whole-body fat mass and waist circumference. There were no significant differences between HIIT and MICT for any body composition measure, but HIIT required ~40% less training time commitment. Running training displayed large effects on whole-body fat mass for both HIIT and MICT (standardized mean difference -0.82 and -0.85, respectively), but cycling training did not induce fat loss.

Conclusions:
Short-term moderate-intensity to high-intensity exercise training can induce modest body composition improvements in overweight and obese individuals without accompanying body-weight changes. HIIT and MICT show similar effectiveness across all body composition measures suggesting that HIIT may be a time-efficient component of weight management programs.

Full text

Etiology and Pathophysiology
The effects of high-intensity interval training vs.
moderate-intensity continuous training on body
composition in overweight and obese adults: a
systematic review and meta-analysis
M. Wewege, R. van den Berg, R. E. Ward and A. Keech
Department of Exercise Physiology, School of
Medical Sciences, University of New South
Wales, Sydney, Australia
Received 29 January 2017; revised 12
February 2017; accepted 12 February 2017
Address for correspondence: M. Wewege,
Department of Exercise Physiology, School of
Medical Sciences, University of New South
Wales, Sydney, NSW 2052, Australia.
E-mail: m.wewege@unsw.edu.au
Summary
Objective: The objective of this study is to compare the effects of high-intensity
interval training (HIIT) and moderate-intensity continuous training (MICT) for
improvements in body composition in overweight and obese adults.
Methods: Trials comparing HIIT and MICT in overweight or obese participants
aged 18–45 years were included. Direct measures (e.g. whole-body fat mass) and
indirect measures (e.g. waist circumference) were examined.
Results: From 1,334 articles initially screened, 13 were included. Studies aver-
aged 10 weeks × 3 sessions per week training. Both HIIT and MICT elicited signif-
icant (p<0.05) reductions in whole-body fat mass and waist circumference. There
were no significant differences between HIIT and MICT for any body composition
measure, but HIIT required ~40% less training time commitment. Running training
displayed large effects on whole-body fat mass for both HIIT and MICT (standard-
ized mean difference 0.82 and 0.85, respectively), but cycling training did not
induce fat loss.
Conclusions: Short-term moderate-intensity to high-intensity exercise training
can induce modest body composition improvements in overweight and obese
individuals without accompanying body-weight changes. HIIT and MICT show
similar effectiveness across all body composition measures suggesting that HIIT
may be a time-efficient component of weight management programs.
Keywords: Exercise, high-intensity interval training, obesity.
Introduction
Obesity, or more specifically the accumulation of excess
body fat, is a significant and rapidly increasing global health
issue. More than 39% of adults were considered overweight
(body mass index [BMI] >25 kg m
2
) and 13% considered
obese (BMI >30 kg m
2
) in 2014, and the prevalence of
overweight and obesity has doubled globally (1,2) since
1980 (1,2). Being overweight or obese is a major risk factor
for cardiovascular and metabolic disorders, in particular,
atherosclerosis, type II diabetes and metabolic syndrome
(3), and increases risk of all-cause mortality (4). In particu-
lar, central adiposity, which specifically relates to adipose
tissue deposited around the trunk and includes the visceral
fat around the central organs, induces a range of negative
adaptations in cardiovascular structure and function, which
magnifies risk of chronic illness and mortality (5–7).
The benefits of physical activity for weight control, reduc-
ing central adiposity and managing obesity are well docu-
mented (8–12). A key recent finding from a meta-analysis
of 117 studies reveals physical activity to be mildly effective
for reducing total body weight (although less effective than
hypocaloric diet) but has a larger effect in reducing visceral
adiposity (10). However, the optimal ‘dose–response’
characteristics of exercise on body composition are still to
be determined, specifically in relation to regional-specific
obesity reviews doi: 10.1111/obr.12532
© 2017 World Obesity Federation Obesity Reviews

changes in central adiposity and visceral fat levels. Tradi-
tional endurance training methods for weight control tended
to focus on longer-duration sessions involving moderate-
intensity exercise performed continuously without rest,
often termed moderate-intensity continuoustraining (MICT).
In recent times, high-intensity interval training (HIIT),
referring to alternating short bursts of high-intensity exercise
and recovery periods, has become a popular alternative
primarily because of its time efficiency, because lack of time
is a commonly cited barrier to exercise participation (13).
There is robust and growing evidence that HIIT may elicit
superior benefits than MICT across a range of health
markers in both healthy and chronic illness populations.
Recent meta-analyses have reported that HIIT induces greater
improvements in cardiorespiratory fitness than MICT in
healthy, young to middle-aged adults (14,15) and in patients
with coronary artery disease and cardio-metabolic disorders
(16–18). Separate meta-analyses specifically focusing on
sprint interval training, a lower-volume variant of HIIT,
which involves repeated intervals of very high-intensity with
a maximum duration of 30 s, also report a moderate-to-large
effect on cardiorespiratory fitness in comparison with no-
exercise healthy control participants (19,20). HIIT also
appears to be superior to MICT for improving markers of
vascular function in patients with cardiovascular or metabolic
disorders (see review (21)), and HIIT is effective for improving
fasting glucose levels and reducing blood pressure in
overweight or obese populations (see review (22)).
Despite clear evidence for the positive adaptations follow-
ing HIIT compared with MICT with regard to aerobic
fitness and vascular function, it is still unclear which form
of training is most effective for weight control, overall fat
loss or central adiposity. Recent studies have analysed the
comparative effectiveness of HIIT and MICT on body fat loss
in overweight populations with varying findings (23–35),
but a systematic review is yet to be conducted. The aim of
this systematic review and meta-analysis was to compare
the effectiveness of HIIT and MICT on body weight and
body composition outcomes in healthy but overweight or
obese adults. Secondary aims were to examine outcomes
specific to central adiposity and factors in exercise
programming which influence effectiveness on body
composition measures.
Methods
Literature search strategy
The review was conducted in accordance with the Preferred
Reporting Items for Systematic Reviews and Meta-analyses
statement guidelines (36). A systematic search of electronic
databases was conducted up to 1 September 2016,
including MEDLINE, Scopus, Embase, SportDiscus, Web
of Science, CINAHL and Physiotherapy Evidence Database
(PEDro). The search strategy comprised key phrases ‘high
intensity interval’or ‘high intensity intermittent’or ‘sprint
interval’to identify relevant trials. The search strategy was
limited to human subjects and randomized controlled trials
where the option was available. The reference lists of the
included studies were also examined for any new references
not found during the initial electronic search. Two reviewers
independently appraised papers (R. V. and M. W.); a third
reviewer (A. K.) was consulted to resolve disputes. Study
quality was assessed using a modified PEDro score (37)
(Table S1).
Inclusion and exclusion criteria
Type of study
The search included studies involving randomized
controlled trial or matched controlled trial designs, written
in English. Uncontrolled, cross-sectional and animal studies
were excluded from analysis.
Type of participants
The review covered studies that included apparently healthy
overweight or obese individuals with mean age between 18
and 45 years. Overweight was defined as a BMI greater
than 25; obese was defined as BMI greater than 30.
Participants were not diagnosed with any other medical
comorbidities such as coronary artery disease or diabetes.
Type of interventions
Training programs were a minimum duration of 4 weeks.
Participants were allocated to a HIIT group or a matched
comparator group that undertook MICT. HIIT programs
involved interval durations of up to 4 min, with the intensity
classified as being greater than 85% of heart rate maximum
(HRmax) or a surrogate physiological index, namely, 80%
maximal aerobic capacity, or a rating of perceived exertion
of 17 (38). MICT programs included continuous aerobic
exercise of intensity 60–75% of HRmax (or 50–65%
maximal aerobic capacity; 12–15 rating of perceived
exertion). Studies were excluded if training was combined
with other forms of exercise training (e.g. resistance
training), but studies that involved a supplementary
nutritional intervention for both trial groups were included.
Outcome measures
The primary outcomes were commonly assessed direct mea-
sures of body composition, including whole-body fat levels,
whole-body lean mass or regional-specific fat measures such
as trunk fat (mass or area) and visceral fat mass (area or
volume). In instances where whole-body fat percentage
was reported, these values were recalculated as whole-body
fat mass (BFkg) using the pre-training and post-training
body mass values. Secondary outcomes were commonly
assessed indirect or surrogate measures of body
2Exercise for improving body composition M. Wewege et al. obesity reviews
© 2017 World Obesity FederationObesity Reviews

composition, which include body mass, BMI and waist
circumference. Whole-body or regional fat mass measures
were drawn from studies that applied hydrostatic weighing,
ultrasound, dual-energy X-ray absorptiometry (DXA), bio-
electrical impedance analysis (BIA), computed tomography
(CT) or magnetic resonance imaging (MRI) (39,40).
Measures of body fat drawn from skinfold measures or air
displacement plethysmography (Bod Pod) were not included
because of validity and/or reliability concerns (41–43).
Data synthesis
Two reviewers (M. W. and R. V.) extracted data in duplicate
and cross-checked results. Outcomes for body composition
were extracted and archived in a database for analysis,
including baseline and post-intervention mean ± standard
deviation values, and mean difference (MD) and 95%
confidence intervals were reported. If not reported, the
MD between pre-intervention and post-intervention was
calculated by subtracting baseline from post-intervention
values. Standardized mean difference (SMD) was calculated
as a percentage change from baseline and was applied when
different methods were used to establish the same outcome
measure. If not reported, 95% confidence intervals and
standard deviations for overall treatment effects were
calculated using Review Manager (REVMAN) 5.3 (Nordic
Cochrane, Denmark). Authors of included studies were
contacted for missing values where required.
Statistical analysis
Between-group meta-analyses were completed for continu-
ous data by using the change in the mean and standard
deviation of outcome measures as outlined previously. A
random effects inverse variance analysis was used with the
effects measure of SMD for BFkg, lean mass, trunk fat, body
mass and BMI measures, and MD for waist circumference.
Heterogeneity was quantified using the Cochrane Qtest
and Higgins I
2
. Egger plots were provided to assess the risk
of publication bias. Independent sample t-tests were
conducted to assess differences between HIIT and MICT
interventions in training hours per week during the inter-
ventions. Within-group meta-analyses were completed for
continuous data using the baseline and post-intervention
values for each intervention. Random effects inverse vari-
ance analysis was also used with the same effects measures
as above. A sub-analysis was also conducted for BFkg and
body mass with studies pooled according to exercise modal-
ity (treadmill running vs. cycle ergometer). Level of signifi-
cance was set at p<0.05 and 95% confidence intervals.
Magnitude of effect was categorized as large (SMD >0.8),
medium (SMD 0.5–0.8), small (SMD 0.2–0.5) or trivial
(SMD <0.2) (44,45). Statistical analysis was conducted
using SPSS 22.0 (IBM Corp., Armonk, NY, USA), and figures
were produced using REVMAN 5.3.
Results
Studies included in the review
The search strategy identified 4,228 articles from electronic
databases, and one additional record was identified through
other means. Following removal of duplicates, 1,334
articles were initially screened via title and abstract, and
63 were identified as potentially relevant. Full-text examina-
tion further excluded 50 studies, leaving 13 studies for
inclusion in this analysis (Fig. 1).
Thirteen studies examined body composition outcomes
in 424 overweight and obese adults (50% male); 216
completed a HIIT intervention (50% male; mean
age = 32.3 years; mean BMI = 29.8), and 208 completed
a MICT intervention (50% male; mean age = 31.5 years;
mean BMI = 29.5). Participant demographics are outlined
in Table 1, and intervention characteristics are outlined in
Table 2. Interventions were conducted for 10.4 ± 3.1 weeks
(range 5–16 weeks), and the mode of exercise was
matched for HIIT and MICT interventions in all studies.
Cycle ergometer was the most common modality (seven
studies (24,25,27,30,32,34,35)), followed by treadmill
running (six studies (23,26,28,29,31,33)).
The HIIT intervention participants trained 3.3 ± 0.7 d per
week, for a total of 95 ± 46 min per week. MICT partici-
pants trained 3.7 ± 0.9 d per week, for a total of
158 ± 46 min per week. Duration of training per week
was significantly lower for the HIIT intervention compared
with MICT (p= 0.003). Of the 13 studies, seven applied
protocols matched for workload or energy expenditure.
All data for whole-body fat levels (N= 11 studies) were
drawn from studies applying either DXA (seven studies
(24,25,28,31,32,34,35)) or BIA (four studies (23,26,30,33)).
Six studies reported whole-body lean mass (four studies
using DXA (27,28,34,35) and two studies using BIA
(26,30)), four studies reported trunk fat values (three studies
using DXA (25,27,35) and one using BIA (30)) and one CT
study provided visceral fat as a measure (33).
Within-group effects
Within-group analyses are summarized in Table 3. Forest
plots for BFkg and body mass are shown in Figs 2a and
3a, respectively. Baseline and follow-up values were
unavailable from one study (27) for trunk fat, and one
study (26) for lean mass, which were not included
in the within-group analyses. The magnitude of change
that these studies reported was used in the between-
group analyses.
Exercise for improving body composition M. Wewege et al. 3obesity reviews
© 2017 World Obesity Federation Obesity Reviews

Pooled analyses of the individual interventions deter-
mined that both HIIT and MICT protocols resulted in
statistically significant reductions in BFkg (SMD: 0.44
and 0.5 for HIIT and MICT, respectively; MD: 1.7 and
2.1 kg for HIIT and MICT, respectively) and waist
circumference (MD: 3 cm for HIIT and MICT each).
There was no significant effect of HIIT or MICT on body
mass (SMD: 0.17 and 0.18 for HIIT and MICT,
respectively; MD: 2 and 1.9 kg for HIIT and MICT,
respectively), lean mass or trunk fat measures. Sub-analysis
of the effect of exercise mode (running or cycling) showed a
large effect following a HIIT and MICT running protocol
for BFkg (SMD: 0.82 and 0.85 for HIIT and MICT,
respectively; MD: 2.6 kg for HIIT and MICT each) and
small effect on body mass (SMD: 0.31 and 0.3 for HIIT
and MICT, respectively) (Table 3). No significant effects
were identified for HIIT or MICT cycle ergometer protocols
on these measures.
Between-group effects
Comparisons of HIIT and MICT interventions revealed no
significant differences in their effects on any measure of
body composition (Table 3). Forest plots for BFkg and body
mass are shown in Figs 2b and 3b, respectively. Sub-analysis
conducted for exercise mode (cycle ergometer or treadmill
running) similarly did not detect any between-group effects.
No heterogeneity was detected in any of these analyses.
Quality assessment
A PEDro (37) assessment determined that the quality of
studies in this analysis was moderate (mean score = 4.6 ± 1.3;
range 2 to 7; Table S1). Common limitations were related to
the allocation of subjects and high dropout rates. All studies
except one (28) (92%) randomly allocated subjects to inter-
vention groups, but only one (24) study (8%) used a
concealed method of allocation. No study blinded subjects
or therapists, but the authors acknowledge the difficulty of
applying this in training studies. However, only three stud-
ies (25,26,33) blinded a study assessor, which presents a
limitation due to potential bias. The overall dropout rate
(from N= 10 studies reporting data) was 18.5% (HIIT:
16%, MICT: 20%, p= 0.074; running studies: 23%,
cycling studies: 15%, p= 0.12). Eight studies (62%) lost
more than 15% of participants to follow-up (Table S1);
Figure 1 Preferred Reporting Items for Systematic Reviews and Meta-analyses flow diagram for study selection.
4Exercise for improving body composition M. Wewege et al. obesity reviews
© 2017 World Obesity FederationObesity Reviews

only Fisher (24) and Sawyer (35) applied an intention-to-
treat analysis to overcome this issue. From four running
studies that provided data (N= 129), 17 participants
(13%) reported an adverse event (eight HIIT participants
and nine MICT participants). Only one cycling study
provided adverse events data, with only one participant
adverse event report provided. No studies reported acute
injuries from either training protocol, with all adverse
events reported as chronic flare-ups or intolerance.
Heterogeneity and publication bias
Moderate heterogeneity was detected in two analyses: the
pooled within-group analysis of HIIT interventions for
changes to BFkg (I
2
=48%) and also the BFkg sub-analysis
for HIIT running protocols (I
2
=69%). This heterogeneity is
due to the strong results of two studies that both applied
running protocols (31,33). Egger plots for all analyses
determined no indication of publication bias.
Discussion
To our knowledge, this is the first review to directly com-
pare HIIT and MICT exercise protocols for changes in body
composition focusing on overweight and obese individuals.
Our results revealed, firstly, short-term aerobic exercise
training of at least moderate-intensity can induce significant
improvements in BFkg and waist circumference, even in the
absence of changes in body weight. Secondly, both HIIT and
MICT appear to be similarly effective on these measures, de-
spite HIIT training requiring ~40% less time commitment.
Thirdly, training programs that involve running appear
especially effective for inducing changes in these body com-
position measures, while cycling programs are not effective.
Each of these findings has major implications for optimizing
weight management interventions.
The primary finding is ~10 weeks of high-intensity or
moderate-intensity exercise training can reduce BFkg by
~2 kg and waist circumference by ~3 cm in the absence of
body mass changes. These values indicate a modest im-
provement in body composition from short-term exercise
training, with body fat mass decreasing by ~6% from initial
levels. It should be noted, however, that the magnitude of
these changes is within the error for repeated measurement
of whole-body fat levels drawn from DXA and BIA if test
conditions have not been well-controlled across sessions
(46–49), and for waist circumference (50) drawn from over-
weight or obese populations, so caution must be applied
when interpreting this finding.
Exercise has consistently been reported to be relatively
ineffective for managing overweight or obesity when not
combined with a dietary intervention, based primarily on
studies that only analysed body mass or BMI (8,51). While
the evidence we present is not definitive, it does add to other
Table 1 Participant demographics for included studies and outcome measure instruments applied
Study Year HIIT MICT Outcome
measure
instrument(s)
Sample size Gender Age (years) Weight (kg) BMI (kg m
2
) Sample size Gender Age (years) Weight (kg) BMI (kg m
2
)
Ahmadizad 2015 10 10 M 25.0 ± 1.0 83.9 ± 3.8 27.6 ± 1.9 10 10 M 25.0 ± 1.0 84.9 ± 4.5 27.6 ± 1.9 BIA
Fisher 2015 15 15 M 20.0 ± 1.5 94.3 ± 12.1 30.0 ± 3.1 13 13 M 20.0 ± 1.5 89.7 ± 15.8 29.0 ± 3.4 DXA
Keatingi 2014 13 3 M; 10 F 41.8 ± 2.7 76.1 28.2 ± 0.5 13 2 M; 10 F 44.1 ± 1.9 80.7 ± 1.5 28.5 ± 0.6 DXA
Kemmleri 2014 33 33 M 43.9 ± 5.0 91.5 ± 14.0 NA 32 32 M 42.9 ± 5.1 89.5 ± 12.3 NA BIA
Kongi 2016 13 13 F 21.5 ± 4.0 69.1 ± 9.5 25.8 ± 2.6 13 13 F 20.5 ± 1.9 67.5 ± 7.3 25.5 ± 2.1 DXA
Martinsi 2015 13 4 M; 9 F 33.9 ± 7.8 97.5 ± 17.0 33.2 ± 3.5 13 5 M; 8 F 33.0 ± 9.9 101.1 ± 14.1 33.3 ± 2.4 DXA
Nyboi 2010 8 8 M 37.0 ± 3.0 96.3 ± 3.8 NA 9 9 M 31.0 ± 2.0 85.8 ± 5.5 NA DXA
Sawyeri 2016 9 5 M; 4 F 35.6 ± 8.9 112.7 ± 26.6 37.4 ± 6.2 9 4 M; 5 F 34.8 ± 7.7 99.7 ± 10.9 34.5 ± 3.2 DXA
Schjervei 2008 14 3 M; 11 F 46.9 ± 2.2 114.0 ± 5.7 36.6 ± 1.2 13 3 M; 10 F 44.4 ± 2.1 104.1 ± 4.5 36.7 ± 1.4 DXA
Shepherdi 2015 46 15 M; 31 F 42.0 ± 11.0 78.8 ± 18.3 27.7 ± 5.0 44 15 M; 29 F 43.0 ± 11.0 77.5 ± 15.8 27.7 ± 4.6 BIA
Sijiei 2012 17 17 F 19.8 ± 1.0 73.7 ± 7.5 27.72 ± 1.88 16 16 F 19.3 ± 0.7 74.2 ± 9.0 28.32 ± 1.96 DXA
Simi 2015 10 10 M 31.0 ± 8.0 87.4 ± 7.7 27.4 ± 1.6 10 10 M 31.0 ± 8.0 86.5 ± 8.6 27.2 ± 1.5 DXA
Zhangi 2015 12 12 F 21.0 ± 1.0 66.4 ± 9.3 25.8 ± 2.7 12 12 F 20.6 ± 1.2 64.8 ± 6.1 26.0 ± 1.6 BIA (BFkg);
CT (visceral)
BIA, bioelectrical impedance analysis; BMI, body mass index; CT, computed tomography; DXA, dual-energy X-ray absorptiometry; F, female; HIIT, high-intensity interval training; M, male; MICT, moderate-
intensity continuous training; NA, not available.
Exercise for improving body composition M. Wewege et al. 5obesity reviews
© 2017 World Obesity Federation Obesity Reviews

Table 2 Program characteristics for HIIT and MICT interventions
Study Duration
(weeks)
Exercise
modality
HIIT MICT
Exercise intensity
(% max)
(interval : rest)
Frequency
(days/week)
Exercise time
per week (min)
Attendance rate,
dropouts and adverse
events
Exercise
intensity
(% max)
Frequency
(days/week)
Exercise
time per
week
(min)
Attendance rate, dropouts
(%) and adverse events
Ahmadizad 6 Running 90% VO
2
max 3 85.5 NR 50–60%
VO
2
max
3 150 NR
Fisher 6 Cycle 85% max AP 3 60 Attendance NR;
Dropouts = 2 (13%);
Adverse NR
55–65%
VO
2
max
5 262.5 Attendance NR;
Dropouts = 3 (23%);
Adverse NR
Keating 12 Cycle 30–45 s 120% VO
2
max: 120–
180 s 30 W
3 66 Attendance = 96%;
Dropouts = 2 (15%);
Adverse = 0
50–65%
VO
2
max
3 126 Attendance = 92%;
Dropouts = 2 (15%);
Adverse = 1 (8%)
Kemmler 16 Running 92–110% IAT-HR 90 s12 min: 1-
to 3-min rest
4 212 Attendance = 83%;
Dropouts = 7 (18%);
Adverse = 3 (9%)
70–82.5%
IAT-HR
4 228 Attendance = 82%;
Dropouts = 9 (21%);
Adverse = 4 (13%)
Kong 5 Cycle 8-s sprint: 12-s rest for 20 min 4 80 Attendance = 100%;
Dropouts = 2 (13%);
Adverse NR
60%
VO
2
max
4 160 Attendance NR;
Dropouts = 3 (19%);
Adverse NR
Martins 12 Cycle 85–90% HRmax 3 60 Attendance = 100%;
Dropouts = 3 (19%);
Adverse NR
70%
HRmax
3 96 Attendance = 100%;
Dropouts = 1 (7%);
Adverse NR
Nybo 12 Running 5 × 2 min >95% HRmax 2 40 Attendance = 67%;
Dropouts NR;
Adverse = 5 (63%)
80%
HRmax
2.5 150 Attendance = 83%;
Dropouts NR;
Adverse = 2 (22%)
Sawyer 8 Cycle 10 × 1 min 90–95% HRmax 3 87 Attendance = 100%;
Dropouts = 4 (18%);
Adverse NR
70–75%
HRmax
3 120 Attendance = 100%;
Dropouts = 2 (18%);
Adverse NR
Schjerve 12 Running 4 × 4 min 85–95% HRmax: 3-min
rest 50–60% HRmax
375 NR 50–60%
HRmax
3 141 NR
Shepherd 10 Cycle 15–60 s >90% HRmax 3 65.1 Attendance = 83%;
Dropouts = 4 (9%);
Adverse NR
70%
HRmax
5 180 Attendance = 61%;
Dropouts = 8 (9%);
Adverse NR
Sijie 12 Running 5 × 3 min 85% VO
2
max: 3 min
50% VO
2
max
5 135 Attendance NR;
Dropouts = 3 (18%);
Adverse = 0
50%
VO
2
max
5 200 Attendance = NR;
Dropouts = 4 (25%);
Adverse = 0
Sim 12 Cycle 15 s 170% VO
2
max: 60 s 32%
VO
2
max
3 112.5 Attendance = 98%;
Dropouts = 0;
Adverse NR
60%
VO
2
max
3 112.5 Attendance = 97%;
Dropouts = 0;
Adverse NR
Zhang 12 Running 4 × 4 min 85–95% HRmax: 3 min
50–60% HRmax; 7-min rest
4 160 Attendance = 94%
Dropouts = 2 (17%);
Adverse = 0
60–70%
VO
2
max
4 132 Attendance = 90%
Dropouts = 3 (25%);
Adverse = 3 (25%)
Mean 10.4 Mean 3.3 95.2 Mean 3.7 158.3
SD 3.1 SD 0.7 46.3 SD 0.9 43.0
AP, aerobic power; APMHR, age-predicated maximum heart rate; HIIT, high-intensity interval training; HRmax, heart rate maximum; IAT-HR, individual aerobic threshold heart rate; MICT, moderate-intensity con-
tinuous training; NR, not reported; VO
2
max, maximal aerobic capacity; W, watts.
6Exercise for improving body composition M. Wewege et al. obesity reviews
© 2017 World Obesity FederationObesity Reviews

recently published evidence (analysing studies that applied
CT and MRI) that exercise could have utility for fat loss,
particularly visceral fat, even if weight loss is not observed.
A recent meta-analysis of 117 studies (N= 4,815 partici-
pants) (10) reported that, while caloric restriction is more
effective than exercise for weight loss, exercise is more effec-
tive for decreasing visceral fat stores, and using correlation
analysis noted that in the absence of weight loss exercise
training still induces ~6% drop in visceral fat. Similarly, a
separate meta-analysis of 29 aerobic training studies (rang-
ing from 4 to 52 weeks and involving moderate-intensity
to high-intensity exercise training) reported aerobic exercise
was effective (effect size 0.33) for lowering visceral fat
compared with control groups (52). As such, our findings
support the growing view that weight management pro-
gramming for overweight or obese individuals cannot focus
primarily on measures of body weight or BMI and needs to
expand to include direct measures of body fat levels (or, at
the least, the indirect measure of waist circumference) to
give a broader indication in the change to the individual’s
overall risk profile (53).
There was insufficient evidence from studies providing di-
rect measurement of trunk fat or visceral fat from CT/MRI
to support the data from the indirect measure of central
adiposity (waist circumference). All four studies reporting
trunk fat were drawn from DXA or BIA, and no exercise-
induced effect was seen. To date, only one study has com-
pared the effect of HIIT and MICT on visceral fat using a
‘gold-standard’instrument for measurement non-invasively
(CT or MRI) (39). This study (33) using CT scanning found
a significant decrease (19.5%) in abdominal visceral fat area
following 12 weeks of HIIT treadmill running, but no signif-
icant decrease in the MICT group (11.1%). Visceral fat
deposits may have more impact on health than simply excess
total fat accumulation (5,7,53), and visceral fat accumula-
tion is independently associated with health issues such as
hypertension and insulin resistance (54,55). Furthermore,
aerobic exercise appears to be a key factor in reducing vis-
ceral adiposity compared with diet, with significant
reductions occurring without changes to overall body mass
(10,52). It is plausible that regional and whole-body fat re-
duction may occur differently between HIITand MICT exer-
cise regimes, primarily because of mechanistic factors related
to mitochondrial adaptations (56). A greater effect from
HIIT on visceral fat adiposity has been reported in women
with type II diabetes (57). Future studies should endeavour
to include visceral adipose measures using CT and MRI.
The second key finding from this meta-analysis was that
HIIT and MICT induced similar magnitude of changes in
BFkg and waist circumference. Considering HIIT involved
~40% less time commitment than MICT in the studies we
analysed, and also demonstrated a comparable dropout
rate, HIIT may be a time-efficient and sustainable strategy
to induce modest improvements to body composition. Lack
of time is reported to be a strong barrier for many people to
undertake physical activity (13,58), so an intensive exercise
program with less time commitment may provide a suitable
option for individuals trying to improve their body compo-
sition. In addition, there is some preliminary evidence that
participants report HIIT to be at least as enjoyable as
MICT, if not more so (58,59), and preliminary data from
application of HIIT in higher-risk populations such as
patients with coronary artery disease do not show an
increased risk of adverse events occurring (59). In combina-
tion with the lower time commitment to training, these
points suggest that HIIT may have utility as a sustainable,
Table 3 Summary of meta-analyses
Outcome (sub-group) Within-group effects Between-group effects
Studies
(n)
HIIT MICT Heterogeneity
nSMD pvalue nSMD pvalue SMD 95% CI pvalue I
2
pvalue
BFkg 11 180 0.44 0.005 & 178 0.5 0.0005 0.03 0.18, 0.24 0.79 0% 0.97
(mode = run) 5 80 0.82 0.01 &790.85 0.001 0.04 0.36, 0.27 0.78 0% 0.64
(mode = cycle) 6 100 0.17 0.23 99 0.23 0.10 0.09 0.19, 0.30 0.55 0% 0.99
Trunk fat no. 4 69 0.19 0.31 71 0.14 0.43 0.10 0.49, 0.28 0.60 13% 0.33
Lean mass no. 6 118 0.07 0.63 120 0.06 0.7 0.16 0.23, 0.55 0.42 49% 0.08
Mass 13 210 0.17 0.09 205 0.18 0.08 0.09 0.10, 0.28 0.36 0% 0.55
(mode = run) 6 94 0.31 0.04 92 0.3 0.04 0.10 0.19, 0.39 0.50 0% 0.51
(mode = cycle) 7 116 0.06 0.66 113 0.08 0.54 0.09 0.19, 0.37 0.52 7% 0.38
Waist circumference 5 83 3.07*0.03 80 3.04*0.006 0.05* 1.09, 1.00 0.93 0% 0.78
BMI 9 143 0.22 0.06 140 0.32 0.02 0.09 0.15, 0.32 0.46 0% 0.57
Bold indicates significant change (p<0.05); * indicates mean difference instead of standardized mean difference; # indicates data missing from one
study, not included in within-group analysis; & indicates significant heterogeneity.
For between-group effects, a positive SMD value indicates greater magnitude of change for the MICT groups compared with HIIT, except for lean mass
where the positive SMD indicates greater magnitude of change that was recorded for HIIT.
BFkg, body fat mass (kg); CI, confidence intervals; HIIT, high-intensity interval training; MICT, moderate-intensity continuous training; NA, not available;
SMD, standardized mean difference.
Exercise for improving body composition M. Wewege et al. 7obesity reviews
© 2017 World Obesity Federation Obesity Reviews

long-term intervention for many overweight or obese
individuals. More long-term and ‘real-world’studies are
required in this area (60), and it is important to note that
HIIT and MICT can play complementary roles in exercise
prescription for managing obesity and even in combination
with resistance training (52).
It is interesting to note the third finding that changes in
body composition is influenced by the choice of exercise
modality. Our results showed that treadmill running in
either HIIT or MICT resulted in significant decreases in
BFkg and even body mass. The magnitude of the effect on
BFkg was observed to be large (SMD 0.82 for HIIT;
0.85 for MICT), equating to 2.6 kg of fat loss in each
group, or ~10% drop in the BFkg from baseline levels. In
contrast, cycle training did not significantly affect any
measure of body composition. The underlying physiological
basis for these differences is unclear and presents a novel
area for research. There are a number of physiological
Figure 2 (a) Forest plot for within-group effects of HIIT and MICT interventions on body fat (kg) and (b) forest plot for between-group effects of HIIT and
MICT interventions on body fat (kg). CI, confidence interval; HIIT, high-intensity interval training; MICT, moderate-intensity continuous training; SD, stan-
dard deviation. [Colour figure can be viewed at wileyonlinelibrary.com]
8Exercise for improving body composition M. Wewege et al. obesity reviews
© 2017 World Obesity FederationObesity Reviews

differences between running and cycling that could plausi-
bly at least partially explain the finding. These include more
muscle mass recruitment during running for any given
sub-maximal workload relative to maximal capacity (e.g.
%HRmax), leading presumably to greater energy
expenditure (61) although this is not yet clear (62). How-
ever, the suitability of applying running training for obese
individuals needs clarification. The running and cycling
studies showed comparable dropout rates (23% and 15%,
respectively), and while running studies constituted a
Figure 3 (a) Forest plot for within-group effects of HIIT and MICT interventions on body mass and (b) forest plot for between-group effects of HIIT and
MICT interventions on body mass. CI, confidence interval; HIIT, high-intensity interval training; MICT, moderate-intensity continuous training; SD, standard
deviation. [Colour figure can be viewed at wileyonlinelibrary.com]
Exercise for improving body composition M. Wewege et al. 9obesity reviews
© 2017 World Obesity Federation Obesity Reviews

considerable rate of reported adverse events (13% across
four studies that reported this measure), the lack of reported
adverse events data from cycling studies (N= 1) does not al-
low for any meaningful comparison of relative safety profile
at this stage. Therefore, a determination about the overall
safety and sustainability of running protocols applied in
overweight or obese individuals is limited. Future training
studies should endeavour to adequately report adverse
events, especially those that relate directly to the
intervention.
Strengths and limitations
The quality of included studies and the small pooled sample
size (total of 424 adults) present limitations for this analysis,
with studies ranging from 17 to 90 participants. As
determined in the PEDro assessment, the included studies
generally were limited by the lack of assessor blinding and
high dropout rates that were not accounted for with
intention-to-treat analyses. Inadequate reporting of session
attendance, program adherence and adverse events can also
be added to the key areas for improvement in overall study
quality for the future.
The generalizability of the findings are limited by the
relatively modest magnitude of change in whole-body fat
levels (~2 kg) and waist circumference (~3 cm) drawn from
relatively short-term training studies (~10 weeks), which are
within the error of measurement for the instruments applied
(47,48,50). In addition, the evidence of change in central
adiposity is largely limited to an indirect measure (waist
circumference) because of insufficient number of studies
applying gold-standard instruments (CT and MRI) for
assessing regional-specific changes in trunk fat or visceral
fat. Therefore, while we report statistically significant
reductions in body fat and waist circumference from short-
term HIIT and MICT, these findings can only provide
limited guidance towards clinical relevance at this stage.
More long-term studies assessing direct measures of body
fat using well-validated instruments are required to clearly
deduce the effect of exercise on whole-body and regional-
specific body fat changes.
Conclusions and practical implications
Short-term HIIT and MICT exercise both elicit modest im-
provements, and of similar magnitude, in body fat levels
and waist circumference in overweight and obese adults.
Considering HIIT shows similar efficacy to MICT, but with
~40% less time commitment each week, HIIT can be con-
sidered a time-efficient alternative for managing overweight
and obese individuals. Future studies need to analyse the
effectiveness of HIIT and MICT on visceral adiposity,
considering the health implications of central fat deposition.
Acknowledgement
We would like to thank the authors of included studies who
provided us with extra data for this review.
Conflict of interest statement
The authors declare no conflicts of interest and report no
sources of funding for this study.
Supporting information
Additional Supporting Information may be found online in
the supporting information tab for this article. http://dx.doi.
org/10.1111/obr.12532
Table S1. Results for assessment of study quality using a
modified Physiotherapy Evidence Database (PEDro) score.
References
1. Finucane MM, Stevens GA, Cowan MJ et al. National, regional,
and global trends in body-mass index since 1980: systematic analy-
sis of health examination surveys and epidemiological studies with
960 country-years and 9.1 million participants. Lancet (London,
England) 2011; 377: 557–567.
2. Organisation WH. Obesity and overweight: fact sheet 311.
2016.
3. Poirier P, Giles TD, Bray GA et al. Obesity and cardiovascular
disease: pathophysiology, evaluation, and effect of weight loss.
Circulation 2006; 113: 898–918.
4. Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-
cause mortality with overweight and obesity using standard body
mass index categories: a systematic review and meta-analysis.
JAMA 2013; 309:71–82.
5. Bastien M, Poirier P, Lemieux I, Despres JP. Overview of epide-
miology and contribution of obesity to cardiovascular disease. Prog
Cardiovasc Dis 2014; 56: 369–381.
6. Ghandehari H, Le V, Kamal-Bahl S, Bassin SL, Wong ND.
Abdominal obesity and the spectrum of global cardiometabolic
risks in US adults. Int J Obes (Lond) 2009; 33: 239–248.
7. Oliveros E, Somers VK, Sochor O, Goel K, Lopez-Jimenez F.
The concept of normal weight obesity. Prog Cardiovasc Dis 2014;
56: 426–433.
8. Shaw K, Gennat H, O’Rourke P, Del Mar C. Exercise for over-
weight or obesity. Cochrane Database Syst Rev 2006: CD003817.
https://doi.org/10.1002/14651858.CD003817.pub3
9. Ohkawara K, Tanaka S, Miyachi M, Ishikawa-Takata K,
Tabata I. A dose-response relation between aerobic exercise and
visceral fat reduction: systematic review of clinical trials. Int J Obes
(Lond) 2007; 31: 1786–1797.
10. Verheggen RJ, Maessen MF, Green DJ, Hermus AR, Hopman
MT, Thijssen DH. A systematic review and meta-analysis on the
effects of exercise training versus hypocaloric diet: distinct effects
on body weight and visceral adipose tissue. Obes Rev 2016; 17:
664–690.
11. Chin SH, Kahathuduwa CN, Binks M. Physical activity and
obesity: what we know and what we need to know. Obes Rev
2016; 17: 1226–1244.
12. Shook RP. Obesity and energy balance: what is the role of
physical activity? Expert Rev Endocrinol Metab 2016; 11:511–520.
10 Exercise for improving body composition M. Wewege et al. obesity reviews
© 2017 World Obesity FederationObesity Reviews

13. Trost SG, Owen N, Bauman AE, Sallis JF, Brown W. Corre-
lates of adults’participation in physical activity: review and
update. Med Sci Sports Exerc 2002; 34: 1996–2001.
14. Bacon AP, Carter RE, Ogle EA, Joyner MJ. VO
2
max trainabil-
ity and high intensity interval training in humans: a meta-analysis.
PLoS One 2013; 8: e73182.
15. Milanovic Z, Sporis G, Weston M. Effectiveness of high-
intensity interval training (hit) and continuous endurance training
for VO
2
max improvements: a systematic review and meta-analysis
of controlled trials. Sports Med 2015; 45: 1469–1481.
16. Elliott AD, Rajopadhyaya K, Bentley DJ, Beltrame JF,
Aromataris EC. Interval training versus continuous exercise in pa-
tients with coronary artery disease: a meta-analysis. Heart Lung
Circ 2015; 24: 149–157.
17. Liou K, Ho S, Fildes J, Ooi SY. High intensity interval versus
moderate intensity continuous training in patients with coronary
artery disease: a meta-analysis of physiological and clinical param-
eters. Heart Lung Circ 2016; 25: 166–174.
18. Weston KS, Wisløff U, Coombes JS. High-intensity interval
training in patients with lifestyle-induced cardiometabolic disease:
a systematic review and meta-analysis. Br J Sports Med 2013; 48:
1227–1234.
19. Gist NH, Fedewa MV, Dishman RK, Cureton KJ. Sprint inter-
val training effects on aerobic capacity: a systematic review and
meta-analysis. Sports Med 2014; 44: 269–279.
20. Sloth M, Sloth D, Overgaard K, Dalgas U. Effects of sprint in-
terval training on VO
2
max and aerobic exercise performance: a
systematic review and meta-analysis. Scand J Med Sci Sports
2013; 23: e341–e352.
21. Ramos JS, Dalleck LC, Tjonna AE, Beetham KS, Coombes JS.
The impact of high-intensity interval training versus moderate-
intensity continuous training on vascular function: a systematic
review and meta-analysis. Sports Med 2015; 45: 679–692.
22. Batacan RB, Duncan MJ, Dalbo VJ, Tucker PS, Fenning AS.
Effects of high-intensity interval training on cardiometabolic health:
a systematic review and meta-analysis of intervention studies. Br J
Sports Med 2016. https://doi.org/10.1136/bjsports-2015-095841
23. Ahmadizad S, Avansar AS, Ebrahim K, Avandi M,
Ghasemikaram M. The effects of short-term high-intensity interval
training vs. moderate-intensity continuous training on plasma levels
of nesfatin-1 and inflammatory markers. Horm Mol Biol Clin
Invest 2015; 21: 165–173.
24. Fisher G, Brown AW, Brown MMB et al. High intensity
interval- vs moderate intensity-training for improving cardiometa-
bolic health in overweight or obese males: a randomized
controlled trial. PLoS One 2015; 10: e0138853. https://doi.org/
10.1371/journal.pone.0138853.
25. Keating SE, Machan EA, O’Connor HTet al. Continuous exer-
cise but not high intensity interval training improves fat distribution
in overweight adults. JObes2014; 2014. https://doi.org/10.1155/
2014/834865.
26. Kemmler W, Scharf M, Lell M, Petrasek C, Von Stengel S. High
versus moderate intensity running exercise to impact cardiometa-
bolic risk factors: the randomized controlled rush-study. Biomed
Res Int 2014; 2014. https://doi.org/10.1155/2014/843095
27. Martins C, Kazakova I, Ludviksen M et al. High-intensity
interval training and isocaloric moderate-intensity continuous training
result in similar improvements in body composition and fitness in
obese individuals. Int J Sport Nutr Exerc Metab 2015; 26:197–204.
28. Nybo L, Sundstrup E, Jakobsen MD et al. High-intensity train-
ing versus traditional exercise interventions for promoting health.
Med Sci Sports Exerc 2010; 42: 1951–1958.
29. Schjerve IE, Tyldum GA, Tjonna AE et al. Both aerobic endur-
ance and strength training programmes improve cardiovascular
health in obese adults. Clin Sci 2008; 115 : 283–293.
30. Shepherd SO, Wilson OJ, Taylor AS et al. Low-volume high-
intensity interval training in a gym setting improves cardio-
metabolic and psychological health. PLoS One 2015; 10:
e0139056.
31. Sijie T, Hainai Y, Fengying Y, Jianxiong W. High intensity
interval exercise training in overweight young women. J Sports
Med Phys Fitness 2012; 52: 255–262.
32. Sim AY, Wallman KE, Fairchild TJ, Guelfi KJ. Effects of high-
intensity intermittent exercise training on appetite regulation. Med
Sci Sports Exerc 2015; 47: 2441–2449.
33. Zhang H, Tong TK, Qiu W, Wang J, Nie J, He Y. Effect of
high-intensity interval training protocol on abdominal fat reduction
in overweight Chinese women: a randomized controlled trial.
Kinesiology 2015; 47:57–66.
34. Kong ZW, Fan XT, Sun SY, Song LL, Shi QD, Nie J. Compar-
ison of high-intensity interval training and moderate-to-vigorous
continuous training for cardiometabolic health and exercise enjoy-
ment in obese young women: a randomized controlled trial. PLoS
One 2016; 11: e0158589.
35. Sawyer BJ, Tucker WJ, Bhammar DM, Ryder JR, Sweazea KL,
Gaesser GA. Effects of high-intensity interval training and
moderate-intensity continuous training on endothelial function
and cardiometabolic risk markers in obese adults. J Appl Physiol
2016; 121: 279–288.
36. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred
reporting items for systematic reviews and meta-analyses: the
PRISMA statement. Ann Intern Med 2009; 151: 264–269.
37. Maher CG, Sherrington C, Herbert RD, Moseley AM, Elkins
M. Reliability of the PEDro scale for rating quality of randomized
controlled trials. Phys Ther 2003; 83: 713–721.
38. Borg GA. Psychophysical bases of perceived exertion. Med Sci
Sports Exerc 1982; 14: 377–381.
39. Shuster A, Patlas M, Pinthus JH, Mourtzakis M. The clinical
importance of visceral adiposity: a critical review of methods for
visceral adipose tissue analysis. Br J Radiol 2012; 85:1–10.
40. Lee SY, Gallagher D. Assessment methods in human body
composition. Curr Opin Clin Nutr Metab Care 2008; 11: 566–572.
41. Wells JCK, Fewtrell MS. Measuring body composition. Arch
Dis Child 2006; 91: 612–617.
42. Tseh W, Caputo JL, Keefer DJ. Validity and reliability of the
BOD POD(R) S/T tracking system. Int J Sports Med 2010; 31:
704–708.
43. Le Carvennec M, Fagour C, Adenis-Lamarre E, Perlemoine C,
Gin H, Rigalleau V. Body composition of obese subjects by air
displacement plethysmography: the influence of hydration. Obesity
(Silver Spring, Md) 2007; 15:78–84.
44. Borenstein M, Hedges LV, Higgins JP, Rothstein HR. A basic
introduction to fixed-effect and random-effects models for meta-
analysis. Res Synth Meth 2010; 1:97–111.
45. Cohen J. A power primer. Psychol Bull 1992; 112 : 155–159.
46. Nana A, Slater GJ, Stewart AD, Burke LM. Methodology
review: using dual-energy X-ray absorptiometry (DXA) for the
assessment of body composition in athletes and active people. Int
J Sport Nutr Exerc Metab 2015; 25: 198–215.
47. LaForgia J, Dollman J, Dale MJ, Withers RT, Hill AM. Valida-
tion of DXA body composition estimates in obese men and women.
Obesity (Silver Spring, Md). 2009; 17: 821–826.
48. Shafer KJ, Siders WA, Johnson LK, Lukaski HC. Validity of
segmental multiple-frequency bioelectrical impedance analysis to
estimate body composition of adults across a range of body mass
indexes. Nutrition 2009; 25:25–32.
49. Pietrobelli A, Wang Z, Formica C, Heymsfield SB. Dual-energy
X-ray absorptiometry: fat estimation errors due to variation in soft
tissue hydration. Am J Physiol Endocrinol Metab 1998; 274:
E808–EE16.
Exercise for improving body composition M. Wewege et al. 11obesity reviews
© 2017 World Obesity Federation Obesity Reviews

50. Verweij LM, Terwee CB, Proper KI, Hulshof CTJ, van Meche-
len W. Measurement error of waist circumference: gaps in
knowledge. Public Health Nutr 2013; 16: 281–288.
51. Chin SH, Kahathuduwa CN, Binks M. Physical activity and
obesity: what we know and what we need to know. Obes Rev
2016; 17: 1226–1244.
52. Ismail I, Keating SE, Baker MK, Johnson NA. A systematic re-
view and meta-analysis of the effect of aerobic vs resistance exercise
training on visceral fat. Obes Rev 2012; 13:68–91.
53. Ross R, Bradshaw AJ. The future of obesity reduction: beyond
weight loss. Nat Rev Endocrinol 2009; 5: 319–325.
54. Rheaume C, Arsenault BJ, Belanger S et al. Low cardiorespira-
tory fitness levels and elevated blood pressure: what is the
contribution of visceral adiposity? Hypertension (Dallas, Tex :
1979) 2009; 54:91–97.
55. Fujioka S, Matsuzawa Y, Tokunaga K, Tarui S. Contribution of
intra-abdominal fat accumulation to the impairment of glucose and
lipid metabolism in human obesity. Metabolism 1987; 36:54–59.
56. MacInnis MJ, Gibala MJ. Physiological adaptations to interval
training and the role of exercise intensity. J Physiol 2016 https://doi.
org/10.1113/JP273196.
57. Maillard F, Rousset S, Pereira B et al. High-intensity interval
training reduces abdominal fat mass in postmenopausal women
with type 2 diabetes. Diabetes Metab 2016; 42: 433–441.
58. Welch N, McNaughton SA, Hunter W, Hume C, Crawford D.
Is the perception of time pressure a barrier to healthy eating and
physical activity among women? Public Health Nutr 2009; 12:
888–895.
59. Rognmo O, Moholdt T, Bakken H et al. Cardiovascular risk of
high- versus moderate-intensity aerobic exercise in coronary heart
disease patients. Circulation 2012; 126: 1436–1440.
60. Gray SR, Ferguson C, Birch K, Forrest LJ, Gill JM. High-inten-
sity interval training: key data needed to bridge the gap from
laboratory to public health policy. Br J Sports Med 2016; 50:
1231–1232.
61. Millet GP, Vleck VE, Bentley DJ. Physiological differences
between cycling and running. Sports Med 2009; 39: 179–206.
62. Scott CB, Littlefield ND, Chason JD, Bunker MP, Asselin EM.
Differences in oxygen uptake but equivalent energy expenditure
between a brief bout of cycling and running. Nutr Metab 2006; 3:
1–1.
12 Exercise for improving body composition M. Wewege et al. obesity reviews
© 2017 World Obesity FederationObesity Reviews